The answer is 'sort of'.
I've never used the SDA Archive before, but here is what happened when I tried to download the dataset at
http://sda.berkeley.edu/cgi-bin/hsda?harcsda+calabor.
I first saved the data file "caldata.txt" into a local directory. I then saved the SAS definition file 'calsas.txt' into the same directory.
I renamed the file 'calsas.txt' to 'calsas.sas' so that EViews would recognize it as a SAS program file.
I then drag and dropped my new 'calsas.sas' file onto EViews repeatedly while editing my way around problems using a text editor. Here is what happened:
1) EViews errored with the message "Statement 'TITLE' is not currently supported". I fixed this by commenting out the line:
Code: Select all
/*TITLE "2001-2002 California Workforce Survey";*/
2) EViews errored with the message "Proc format option 'LIB' is currently not supported". I fixed this by commenting out the 'LIB' option as follows:
3) EViews errored with the not particularly useful message "Syntax error: found '(' when expecting ';' ". I fixed this by modifying the line:
Code: Select all
DATA IN /*(LABEL="2001-2002 California Workforce Survey")*/;
4) EViews errored with the message "Statement 'IF' is currently not supported. I fixed this by commenting out the block of lines
Code: Select all
/*IF (age = 99) THEN age = .;
IF (gender = 9) THEN gender = .;
IF (hh18 = 99) THEN hh18 = .;
...
5) EViews errored with the message "Proc Datasets is currently not supported". I fixed this by commenting out the block of lines:
Code: Select all
/*PROC DATASETS;
CONTENTS DATA=IN POSITION;
TITLE 'CONTENTS OF ALL VARIABLES';
RUN;*/
6) EViews reported "Access denied to file 'C:\Users\Chris\Deocuments\SDA\ ' by operating system". I fixed this by specifying the correct filename in the FILENAME statement (note that this step would also have to have been done if you were using SAS itself).
Code: Select all
FILENAME INDATA "caldata.txt"; /* Put name of data file between quotes*/
FInally, I dropped in the file one more time and it read everything in.
I have to admit that step 3) was fairly cryptic, but otherwise, it was largely just a matter of commenting out whatever EViews reported that it didn't support.
My guess is that all files from the same site would probably require pretty much the same edits, so if you did this repeatedly it would become fairly straightforward.
Still, it's clearly not as clean as reading datasets from Data Ferret. As I mentioned, I was unaware of the SDA archive before you mentioned it. If enough users are interested in working with it, we might be able to put in some tweaks to make things a bit easier.